Please help how to configure HTTP Request

Hi everybody,

I’m new to N8N. I have on-prem version 1.101.2.

I need to create an HTTP request to create a user. When I run the HTTP Request manually, I am prompted for parameters and after filling it in, the user is created fine.

When, I enter the instruction via chat. So the AI agent correctly asks me for name, username, email and phone number. But after that nothing happens and the user is not created. According to the GUI display, it no longer used the HTTP request to enter my inputs there.

Hey @migo hope all is well, welcome to the community.

Your system prompt says that the tool name should be “CreateUser”, but on the canvas I see it is called “HTTP Request”. I think this could be contributing the agent confusion. Try naming that node with what you tell AI Agent it is named.

Hi Jabbson,

thank you for your help. I inserted the canvas by mistake. It’s the default scheme. Yesterday I figured out where the problem is. I’m using a custom GPT and there is obviously a problem there. The GPT does know what parameters to ask for because they are requested in the HTTP Request, but then it doesn’t use it to do the POST anymore. Yesterday I decided to try the paid OpenAi and it does perform the POST.
Too bad, because otherwise my GPT works great and I’d like to fix this in the future.

MiGo

Glad you found the issue!
Cheers